Females

Interactive Time Series

2016 Analysis

Subplots

Males

Interactive Time Series

2016 Analysis

Subplots

65 Year Old Females

Interactive Time Series

2016 Analysis

Subplots

65 Year Old Males

Interactive Time Series

2016 Analysis

Subplots

Notes

Boxplot

From the time series on previous pages, it appears that Cardiff is a consistent outlier in terms of its population density. The boxplot here appears to confirm this. To formally identify the outlier, utilise an outlier analysis, using Quartile 3 + (1.5 * IQR)

Outlier Analysis

[1] 1409.14

The above output is the resulting threshold of the outlier analysis. (1409.14) is the outlier threshold for the Welsh population densities in 2016. Any local authority with a population density higher than this threshold is formally labelled an outlier and should be removed from Pearson’s correlation coefficient.

  local_authority Population.density mean_year
1         Cardiff            2563.33      2016

As can be seen above, the only local authority with an outlier population density is Cardiff. Cardiff’s population density is over 80% larger than the outlier threshold. Therefore, it can be removed from Pearson’s correlation test.

Pearson’s Correlation Coefficient

2016, Female Correlation Coefficient (Life Expectancy ~ Population Density)

[1] -0.6714913

In removing Cardiff from the analysis, we have identified a strong negative association between life expectancy and population density. The threshold used to assume a strong negative correlation is -0.5. Below, the coefficients of the same analysis for the other demographics within the data are presented.

2016, Male Correlation Coefficient (Life Expectancy ~ Population Density)

[1] -0.6986888

Interestingly, the data for males returns a stronger negative association than in the female data.

2016, 65 Year Old Female Correlation Coefficient (Life Expectancy ~ Population Density)

[1] -0.6327204

For 65 year old females, there is a slightly weaker negative association between life expectancy and population density. However, this value is still well above the 0.5 threshold used to denote a strong association.

2016, 65 Year Old Male Correlation Coefficient (Life Expectancy ~ Population Density)

[1] -0.7643989

Of all the demographics analysed, the 65 year old males’ life expectancy shows the strongest negative association with population density.

Conclusion

“Is Life Expectancy Affected by Population Density in Welsh Authorities?”

The analyses performed upon the data appear to support the conclusion that population density does affect life expectency within the local auhtorities of Wales. Although the strength of this association varies depending upon the specific demographic being analysed, all demographics available to analyse in the data returned strong correlation coefficients. The coefficients observed are only observed when Cardiff is excluded from the Pearson’s tests, due to its diagnosis as an outlier in terms of its population density.

Some important considerations:

  • Population Density is likely to be a compound variable, with a multitude of significant influences such as gdp, public service demand, pollution and so on.
  • It is likely that analyses of differing demographic groups would result in varying strength of correlation.

Data Sources

Life Expectancy Data

source = Data Cymru Correct as 03/11/2019

Population Density Data

source = Stats Wales Correct as 03/11/2019

Version 1 Notes

  • Mean year has been engineered from the source data by removing inconsistent periods and calculating a mean year. The original data used life expectancies aggregated over 3 year periods.
  • Population densites are taken from mid-year estimates.

  • Please visit the source data websites to see their associated metadata.

  • To get in touch with questions or suggestions for amendments, please get in touch by Email